lundi 10 août 2020

Open letter to Dr Piepho English version



I read with interest the letter you sent a few days ago to the editors of New Microbes and New Infections, Volume 37, entitled "‘Influence of conflicts of interest on public positions in the COVID-19 era, the case of Gilead Sciences’ by Roussel and Raoult (2020)."

I am not a biostatistician as you are. Nevertheless, I feel that your comments are not entirely legitimate.  First of all, you criticize the fact that Roussel and Raoult used what you call aggregation, and the reduction to five items of the opinions analyzed by the authors.  This method is very common in social sciences, and a number of opinion surveys asking for the appreciation of a cohort of people propose to classify the answers into a few categories, usually five, ranking from 1 to 5.  There is therefore no reason to refuse the application of this method to the content of the articles analyzed by Roussel and Raoult.  Moreover, the subdivision into more subtle classes of opinion would have been rightly criticized.  Reducing the analysis to one class would have rendered the document, as well as the opinion survey, uninteresting.  If I understood Roussel and Raoult’s paper correctly, the classification into 5 classes of opinion preceded their allocation to Gilead funding.  Obviously, we cannot rule out this possibility.  However, by suggesting it, you also imply that the authors are not entirely honest or not competent.

The second reason you cite, which is, according to your judgment "irredeemable" (sic), is that Roussel and Raoult have not mentioned the documents and the passages they used to classify the doctors’ opinion.  According to French law, this would probably have been prosecuted under the charge of “slanderous denunciation” (with no chance of success since the data concerning the doctors funding are public).  In my opinion, the merit of this paper is that nobody is directly named and anyone can consult official sites to find out the names of these doctors.  At this point, there are two solutions: either Roussel and Raoult are not honest, or they can be trusted. Although you give the impression that you do not share either of these opinions, it seems that the first one appears in the background of your letter, unless you are denying the statistical expertise of the Marseille team. On this point, I think you are wrong.  Too many statistics kill statistics! Sutor, ne ultra crepidam!

There is one way (often used by Jean Piaget to check the quality of his ranking concerning the stage of children psychologic development) to prove the validity of Roussel and Raoult’s opinion: give the titles of the article and the passages used to rank the opinions on the use of hydroxychloroquine for the treatment of Covid-19, to an anonymous researcher (without mentioning of course the Gilead funding) and compare his ranking with that of the authors you criticize.

I hope to hear from you in the very near future.

Yours sincerely,
Philippe POINDRON





I would like to gratefully acknowledge Mrs Dominique Tousten for her help in correcting my English.

Aucun commentaire: